Mobile Sensor Big Data Challenges in Realizing Precision Medicine

Опубликовано 22 июня 2016, 19:21

National Center of Excellence for Mobile Sensor Data-to-Knowledge (MD2K) is one of the 11 national Centers of Excellence funded by the Big Data-to-knowledge (BD2K) program of the National Institutes of Health (NIH). While other centers are addressing various aspects of biomedical big data (e.g., genomics, informatics), MD2K is leading the effort on development of big data methods and tools specific to the unique complexity of mobile sensor data (characterized by noise, data loss, wearability issues, and confounders) in order to convert it into information, knowledge, and, ultimately, action. In particular, MD2K is using five sensor suites to monitor newly abstinent smokers and congestive heart failure patients – AutoSense for cardiorespiratory monitoring, EasySense for contactless monitoring of cardiorespiratory movement and composition, Microsoft Band for monitoring arm movements in eating, smoking, brushing, and driving behaviors, and GPS and smart eyeglasses to capture environmental cues. It is developing methods for early detection of adverse health events (e.g., pinpoint the timing of first smoking lapse or worsening of lung congestion). Next, it is developing time series pattern mining algorithms and interactive visualization tools to discover predictors of adverse health events, and developing online learning algorithms for delivering just-in-time (JIT) adaptive intervention. Microsoft Band is used not only to detect health-related behaviors, but also to deliver JIT interventions. To facilitate development of these methods and their real-time operation, MD2K is developing a standards-based, interoperable, extensible and open-source big data software platform to support both off-line analysis of mobile sensor data at population-scale and online data analysis at individual scale. Finally, it is conducting field studies with 225 smokers and 225 congestive heart failure patients to evaluate the usability, utility, and validity of markers and predictors discovered by MD2K. Big data analytic tools developed by MD2K will be an essential component of President Obama’s Precision Medicine initiative, which will enable the collection, integration, management, visualization, analysis, and interpretation of health data generated by mobile sensors.